--- layout: post title: Logs for the Monero Research Lab Meeting Held on 2019-04-15 summary: Surae work, Sarang work, and miscellaneous tags: [dev diaries, community, crypto, research] author: el00ruobuob / sarang --- # Logs **\** Let's begin our meeting! **\** 1. GREETINGS **\** Hi **\<[-mugatu-]>** o/ **\** Quiet day today... **\** But I suppose we can still move to 2. ROUNDTABLE **\** suraeNoether: care to go first today? **\** sure, my cat is missing and i want to go look for her, so i'm going to make this quick **\** :( **\** Understood **\** CLSAG signatures are fast and small, they are so fast and small that my naive colored-coin approach could support two assets and still be faster and smaller than our present MLSAG scheme **\** i'm not recommending coloring monero, but commenting on overall speed, it's nuts **\** however, as sarang mentioned, there is a key image problem i'm looking into **\** it's possible rectifying them will cost us some of those gains **\** Yeah, I don't think a straightforward LSAG reduction works here **\** in the meantime, i'm handing CLSAG off to sarang for at least 7 days so i can focus on MRL11 **\** But I wonder if a redefinition of the security requirements to accommodate the new linking will be sufficient **\** since we're rounding the corner on that, it's my top priority, and i want to get CLSAG out of sight for a few days **\** righto **\** okay, i'll be back later today, hopefully with suraecat **\** Thanks, and best of luck with your search **\** I completed the building blocks for a simple Lelantus transaction flow (insecure example code in agenda) **\** and am in contact with the paper's author to discuss some privacy aspects of the construction **\** the CLSAG example code has been updated to reflect some changes **\** and, as suraeNoether said, still working on proper formalization, which is trickier than expected **\** The output selection algorithm discussed here still has an open PR from moneromooo that needs eyeballs **\** PR 5389 **\** Hi **\** yo **\** Will lurk mostly. **\** Does anyone else have research to share? **\** Just announcing presence **\** Otherwise we can keep waxing poetic about CLSAG definitions **\** here now **\** :/ **\** Hi suraeNoether **\** sgp\_: **\** bah, silly autocomplete **\** I have these multi user txes going in the background, and I am wondering whether the 'a' values can be reuesd for multiple outputs. **\** Remind me what these values are/ **\** What's the status on M-of-N multisig? **\** The idea is to make 16 actual outs for the "same" logical output, so they get shuffled as new outputs are added. **\** (our notation is often inconsistent) **\** And I don't know whether it's safe to keep those. I assume sharing them with other usesr of the same tx is not good. **\** \ however, as sarang mentioned, there is a key image problem i'm looking into \<= This is referring to CSLAG right? **\** dEBRUYNE: yes **\** there are no such issues with MLSAG **\** All right, thanks for clarifying **\** The problem refers to the fact that trying to reduce CLSAG to LSAG with an aggregated key yields the wrong key image **\** a is the random secret keys generated at proive time to create the pseudoOuts. **\** Hmm ok **\** You asked me to review this earlier, and it completely slipped my mind **\** I'll look for the code snippet you sent in PM **\** to ensure I don't get wrong the terms you're referring to **\** ty **\** sgp\_: did you have something you wished to discuss too? **\** I don't believe so **\** Well, this meeting is turning out to be quite short :D **\** moneromooo: anything specific, aside from the reuse question you posed? **\** (to discuss here, I mean) **\** Not at the moment I think. **\** OK, I suppose we can move right along then **\** to 3. QUESTIONS and 4. ACTION ITEMS **\** While suraeNoether continues working on matching/churn via MRL-0011, I have several things for the week **\** Now that CLSAG reduction to LSAG is proving so problematic, I want to see if definition modifications for the LSAG proofs will suffice for our use case **\** I'll be checking on moneromooo's question shortly (apologies for letting that slip by) **\** as well as more work on Lelantus transaction flows **\** sarang: Have you consulted RandomRun regarding this problem btw? **\** suraeNoether and I have been in contact with him throughout the development process **\** I don't believe this problem has practical effects on CLSAG's security, only in the complexity of the formalization **\** Any other questions or action items on people's minds? **\** I see, so it does not render the scheme infeasible? **\** Heh, depends **\** If we end up not being able to prove secure under proper definitions, that's a bit of a quandry **\** but in the worst case, we decide not to adopt the scheme, and are right back to where we are now **\** True, better safe than sorry I guess :p **\** However, the space and time savings are so compelling that it's worth the effort **\** ~25% space savings and 15% time savings for a 2-2 transaction **\** (in the signature portion) **\** Not bad **\** OK, any last questions before we formally adjourn? **\** Righto, in that case, we are adjourned. Discussion can of course continue **\** Logs will be posted shortly to the GitHub agenda issue